AI helps you reading Science

AI generates interpretation videos

AI extracts and analyses the key points of the paper to generate videos automatically


pub
Go Generating

AI Traceability

AI parses the academic lineage of this thesis


Master Reading Tree
Generate MRT

AI Insight

AI extracts a summary of this paper


Weibo:
We only modeled the statistical dependency between two adjacent pixels of an image which can be characterized by one-step transition probability matrix of Markov chain

Image tampering detection based on stationary distribution of Markov chain

ICIP, pp.2101-2104, (2010)

Cited: 67|Views18
EI WOS SCOPUS

Abstract

In this paper, we propose a passive image tampering detection method based on modeling edge information. We model the edge image of image chroma component as a finite-state Markov chain and extract low dimensional feature vector from its stationary distribution for tampering detection. The support vector machine (SVM) is utilized as class...More

Code:

Data:

0
Introduction
  • A photograph implies the truth of what was happening. people in the digital world sometimes can not trust image media since (maliciously) tampered images are often found in the Internet even published in newspapers.
  • With the development of image editing software such as Adobe Photoshop, digital image can be manipulated and hardly detected by human eyes.
  • If the authors take this issue for granted, it may eventually be harmful for the digital world, especially for the credibility of news coverage.
  • In this paper the authors focus on passive image tampering detection based on supervised learning techniques
Highlights
  • A photograph implies the truth of what was happening
  • We have proposed a low dimension feature vector extraction method for color image tampering detection
  • We modeled the thresholded edge image of image chroma as a Markov chain and considered its stationary distribution as features for tampering detection
  • The experimental results have illustrated that the proposed 9-D feature vector is very effective for tampering detection
  • We only modeled the statistical dependency between two adjacent pixels of an image which can be characterized by one-step transition probability matrix of Markov chain
  • The actual dependency of image is not limited to just two adjacent pixel, in our future work, the dependency among more than two adjacent pixels would be considered for further analysis
Results
  • The only two available public image databases for tampering detection, specially for splicing detection, are provided by DVMM, Columbia University [12].
  • The CASIA tampered image detection evaluation database (CASIA TIDE) v2.0 [14] consists of 7,491 authentic and 5,123 sophisticatedly tampered color images of different sizes varying from 240 × 160 to 900 × 600
  • This database is with larger size and more realistic and challenging tampered images with complex splicing as well as blurring.
Conclusion
  • The authors have proposed a low dimension feature vector extraction method for color image tampering detection.
  • The authors modeled the thresholded edge image of image chroma as a Markov chain and considered its stationary distribution as features for tampering detection.
  • Though the work in this paper is only focus on image tampering detection, the authors are looking forward the proposed method can be useful at other similar forensics task like device source classification and steganalysis
Tables
  • Table1: Experiment results of proposed method
  • Table2: Experiment results of the method proposed in [11]
Download tables as Excel
Funding
  • This work is funded by research grants from the National Fundamental Research Program of China (Grant No.60603011) and the National Laboratory of Pattern Recognition
Reference
  • C. Rey and J.L. Dugelay, “A survey of watermarking algorithms for image authentication,” EURASIP Journal on Ap-
    Google ScholarLocate open access versionFindings
0
Your rating :

No Ratings

Tags
Comments
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn